均值收斂交易策略在金融市場上一直以來都被廣泛的應用，而均值收斂交易策略在配對交易下的表現往往較單一商品佳。然而，均值收斂模型假設分佈為常態分配，但股票報酬分佈一直以來存在厚尾的問題，而Student′s t 分佈擁有厚尾的特性。本文藉由均值收斂模型假設分佈為Student′s t 分佈下可較佳捕捉厚尾的性質，且利用台灣金融市場的交易資料進行回測，並將其結果與均值收斂模型在常態模型下進行比較。;The mean reverting trading strategy has been widely used in the financial market, and the pair trading performance of the mean reverting strategy is often better than a single commodity. However, in empirical studies, we obtain that the density of log returns, but the density of log return usually has fat-tailed problem. The Student′s t distribution has the characteristic of fat tail. In this paper, we propose the reverting model under the Student′s t distribution can capture better fat-tailed property, and use stock data from Taiwan finance market. We show the empirical results in order to compare the performance of two reverting model.
